Landscape and urban Planning 94 (2010) 264-275, Contents lists available at ScienceDirect, Landscape and Urban Planning journal homepage: www. I sevier,.com/locate/landurbplan ELSEVIER The relation between perceived sensory dimensions of urban green space and stress restoration Patrik Grahn*, Ulrika K. Stigsdotter” + Deparment of Wok scence Busines Economics and Enironental Psychology The Swedish Unversity of Agricultural Scenes, POBox 8.525053 Anap, sweden ® frst Landscape Denmerk University of Copenhagen, Rolghedsvej 25, K 1958 FeersbegC Denmark ARTICLE INFO ABSTRACT Research indicatesa relationship between sensory peception of naturalenvivonments and human heath ur hypothesis i that peopl perceive green spaces in terms of certain dimensions, where some dimen sons afe more important ane preferred than others with respect a restoring people from stress. The aims are to: identity and describe the perceived dimensions in nature; identify which dimen sions people in general prefer; identify the dimensions people reporting stress prefer; and identity combination othe dimensions people reporting stress prefer. ‘Atotal of 952 randomly selected informants from nine Swedish cities (representative of the Swedish population) answered a postal questionnaire with pre-coded questions. The questionnaire consisted of three parts: personal data, preferences fr natural qualities ad sef-estimations of heath status, The data were analyzed using factor analysis and regression analyses The results identify and describe eight perceived sensory dimensions. People in general prefer the dimension Serene followed by Space. Nature, Rich in Species, Refuge, Culture, Prospect and Social. The dimensions Refuge and Nature ae most strongly correlated with stes, indicating need to find the most restorative environments, A combination of Refuge. Nature and Rich in Species, and alow or no presence of Socal, could be interpreted asthe most restorative environment for stressed individuals roma ity planning perspective, the results indicate how urban green paces canbe viewed as elements of importance to public mental health However, before the dimensions can be used by practitioners 3¢ ‘tools to promote health through city planning. more cesearchis needed, {© 2009 Elsevier V.Allrights reserve. 1. Introduction learn very litle about the qualities found in the urban green environments, We live in an urbanized world. About 75% of inhabitants in the During the past three decades, more and more research find- ings have pointed to urban green spaces as aresoutce in promoting, public health. It has been suggested that green spaces promote health by restoring mental fatigue (Kaplan, 2001), serving. as 2 resource for physical activities (Bjork et al, 2008), and reduc- ing all-cause and cause-specific mortality (Mitchell and Popham, 2008), However, most studies have concerned the resources that reduce stress levels (eg. Grahn and Stigsdotter, 2003; Nielsen ‘and Hansen, 2007; Ulrich, 2008). Most research studies have com- pared urban environments that have some kind of nature qualities ‘with urban environments that have no nature qualities at all (eg, van den Berg et al, 2002; Hartig et al, 2003; Velarde et al, 2007). From these studies we learn that access to nature in urban environments is better than no access to nature. Thus, we * Corresponding author Tel: +46 40415495: fax +46 404155045010, mai edreter Farrah (P.Grana, URSBe ks 0169.2046/8~ see font matter © 200 Elsevier BY. Alright reserved 449101016) landurbplan 200810012 developed countrieslive in dense urban areas Habitat,2001).Inthe present debate on the dense city contra the sprawled city. the dense city is often considered a suitable direction for future city planning, the idea being that such cities promote sustainable development (Hardy, 2004), There isa need to understand and analyze the qual- ities found in urban green spaces if we are to make sound decisions about which green spaces have potential to be developed into restorative environments (Velarde et al, 2007). Which qualities in green spaces are popular among and important to inhabitants and which ate not so important? In order to understand this, we need to improve our knowledge and understanding of how people experience and perceive urban green spaces. 11. Perceiving urban green spaces Perception is described as the process of attaining awareness and understanding of sensory information (Bell, 1998). This infor- ‘mation is registered by a variety of sensory cells, and passed onGra, UX Sugar Landscepe and Ura Planning 94 (2010) 258-275, 285 to the sensory organs or senses. It is not easy to define how many senses we have, or even to define what a sense is. Some claim there are five senses, others say twelve (Ayres, 1983), and still others Count even more. Tactile sensations and touch involve many sen- sory cells, such as those registering temperature, humidity, sft touch, pressure, the feeling of texture, etc. Our body mavernents give us information about speed and direction, elevation and the resistance our bodies meet when moving about (Ayres, 1983: Bell 1999), Some senses can be defined as distance senses (Vision, hear- ing and, to some extent, smell), while the other senses are nearness senses (Bell, 1999), However, we rarely use senses in isolation; on the contrary, they most often work in concert (Bundy et al, 2002), The brain interprets the information from all our senses together ‘with our memories of earlier episodes. which gives us afl experi- fence and understanding of our environment (Bell, 1989; Bundy et al, 2002). 112, Sensations, heath and well-being Vision is a sense of utmost importance when visiting urban green spaces, However, other senses are also of great impor- fance, not least concerning the connection between use of urban. green spaces and people's health and well-being. In her program, designed to treat and rehabilitate handicapped children, Ayres (1983) found that information from tactile sensations appears to be basic to the individual's sensory integration, leading to cognitive and sensory-motor development, Moreover, research results indi= ‘ate that information derived from our senses is coded and stored ‘through three different processes: subsymbolic, symbolic imagery, and symbolic verbal (Bucci, 2003). Subsymbolic processing aper- ates in the sensory, motor. and somatic modalities. it helps the soccer player understand how to handle the ball inthe field, andthe ‘wine taster judge the quality of a wine. These knowledge processes ‘occur in specifi sensory-somatic modalities rather than in system- atic cognitive ways: Informations processed via our muscles, inner organs, etc. (Bucci, 2003). In contrast to subsymbolic processing, symbols may be images (eg. a visual picture in a person's mind) words (eg. verbal concepts and interpretations: Bucci, 2003), These three systems have different contents and different prin- ciples of organizing and storing information in our body and brain, However, they are connected by referential links, withthe help of symbolic images, which enable us to symbolize and verbalize out emotional experience and also to understand others (Bucci, 2003) Sometimes the three systems have a harder time connecting—for instance, when people are feeling poorly. Here, sensations and images from the environment can work as a catalyst, mediat- ing information between the three systems, which is of utmost {importance to mental restoration (Bucei, 2003),Searles (1960) also pointed out that signals from nature act as a catalyst, sparking Creative processes that are important to restoration. Complicated relations may be too much to handle, Most complex. are our rela- tions to other people, and the simplest relations are those between, inanimate objects, such as stones, and us. Plants and animals fall, somewhere in between. According to Searles, being able ro mas {er these relationships helps us to recover from crises (Ottosson and Grabn, 2008; Searles, 1960), According to these theories, urban sgreen spaces rich in sensations with no or low demands. which are processed subsymbolically as well as through processing at the symbolic imagery and symbolic verbal levels.may offer possibilities for restoring people's health and well-being If we are to perceive and act appropriately and quickly in the environment, we cannot pay attention to all information coming, {vom our senses. To avoid chaos when making decisions in relation, to everyday rap activities, we need to find order and hierarchies. ‘One of the theories of haw we discern order from the perceived ‘cues presented to us is called Gestalt theory (Perls etal, 1970; Bell 1999), which proposes that separate figures of wholeness Gestalts) stand out from the rest of the environment, Ehrenzweig (2000) efined this conscious, rapid type of perception using Gestalts as “surface perception”. According to Ehrenzweig (2000), sensory information can also be perceived and stored unconsciously in the form of more undifferentiated information, that is, not Gestalts, He defined this type of perception as “depth perception”. Depth perception is needed to get a sense of the “true world”, and it {s normally fluently integrated with surface perception. However, Ehrenzweig also claimed that dissociation between depth percep tion and surface perception (caused by, eg. stress) could cause ‘mental lless, This dissociation can be broken if we have the oppor= tunity and time to make contact with depth perception—through ‘unconscious scanning. Successful symbol formation depends on a fusion between the inner and outer worlds. Hence, our unconscious depth perception can provide us with symbols, helping ustorestore and deepen our sense of reality by helping us find a hidden order in reality (Ehrenzweig, 2000), Although Bucci (2003) did not talk about depth perception, this isin line with her findings, showing. that symbolic images mediate health processes asa referential link between information coded and stored through subsymbolic, sym= bolic imagery and symbolic verbal processes when the systems are dissociated, 1.3, Order and hierarchies Gestalt theory has been used as a basis for claiming that peo- ple find order and hierarchies in the surrounding environment (Bell, 1999). Another theory is related co the ecological approach, to perception, which suggests that we inherently and through con- ditioning look for certain cues or characteristics nour environment that afford us utilities (Gibson, 1975), where pleasure and beauty also constitute a kind of utility that offers us happiness. One exam- ples that humans have always been forced to find shelter from the ‘weather, dangerous animals and enemies. Given thisneed, we have learned to detect what characteristics of the landscape can afford us shelter and hiding places. Moreover, Stern (1985, 1993) claimed, that, even as small infants, we perceive, understand and evaluate the environment by using our cognition and emotions simultane- ously, so-called “vitality affects”, which constitute a pre-linguistic language. Parents and infants use this language to communicate starting rom the infant's first day of life outside the womb. In this way, characteristics in the environment are immediately given a cognitive and emotional label Based on the above, we could say that people interpret the environment in relation to certain classes, where some are more important than others, in that they are varyingly good or bad (Bell, 1999). These classes can be expressed as Gestats or affor- dances, which we look for inherently or through conditioning Moreover, we need peaceful, mote undifferentiated areas where we have a chance to make contact with our basic senses and more deeply stored subsymbolic information. Preference is a concept that assumes a real or imagined “choice” between alternatives and the possiblity of rank ordering these alternatives. The choice is based on the happiness, satisfaction, gratification, enjoyment and utility they provide, More generally, it can also be seen as a source of motivation (Stanford Encyclopedia on Philosophy, 2008) This means that we cannot isolate perception of urban green spaces to vision, especially when we talk about well-being and health, When we are seeking information about the connections between experienced characteristics in urban green spaces and health, we must see perception as involving all our senses. But is it possible to make the general assumption that people prefer cer= tain experienced qualities ~ including scents, touch and sounds ~ ‘more than others, and especially when they are feeling poorly? In266 Gra UK Supt Lendscape and Urban inning 8 (2010) 268-275 order to test that idea, we need a way of dividing the experienced ‘qualities of urban green spaces into certain classes. 14, Classification of perceived sensory environmental information In the 1960s, planners tried to develop techniques for evaluat ing green environments in the countryside. Most often they used ‘expert judgments (Clark 1968), and criteria such as shape. coor, scale and texture were often used. These routines live on, and eval- uations of landscapes are still connected t0 obvious, measurable criteria, such as size, topography, distance and color (Gyllin and Grahn, 2005), However, such evaluations were also criticized early ‘on for the more or less arbitrary variables being judged—a user perspective was considered necessary (Penning-Rowsell, 1973). A ‘roup of researchers began using preference studies, most often by showing photo-slides. in order to establish a foundation for evaluations of green spaces (Shafer etal, 1969). ince then, photo- based studies have been the predominant method in preference studies on large-scale landscapes in the countryside and in urban ‘environments (Gyilin and Grahn, 2005), Considerably fewer stud ies have taken into account senses other than vision (Gyllin and Grabn, 2005). However, Feste and Oterholm (1973) and Oraug, et al (1974), {ound that people's experiences of landscape dimensions include values such as wildness and culture, which is connected to, eg. history. And Appleton (1975) proposed that people have inher ‘ent reflexes causing them to seek optimal places inthe landscape, ‘which have been important for human survival during our pre- history, More recently, several studies have found that experienced ‘qualities in green spaces can be divided into certain characteris tics: most often six to nine (e.g, Grabn, 1991; Grahn and Sorte, 41985; Kytta and Kahila, 2005; Maikov et al, 2008: Stahle, 2005; Tyrvéinen et al, 2007; van Herzele and Wiedemann, 2003), These results show that people's reasons for going out into a landscape are rarely connected to the color or shape of the recreational sit. ‘On the contrary, visitors seek out experiences that they associate ‘with nature and urban parks. Iis likely that the perception of land scapes involves inherent reflexes as well as experiences connected to expectations of, eg. finding different species oF signs of history and culture. The most important finding from these studies is that visitors can identify all characteristics rather clearly and strongly in some green spaces, while other spaces may both lack many characteristics totally and show only weak signs of the remaining characteristics ‘The above-mentioned studies have been carried out in differ- ‘ent ways, concerning establishing the different characteristics: The study by Grahn and Sorte (1985) employed a postal questionnaire sent fo every registered organization in nine Swedish cities. The study by Grahn (1991) was further developed in Berggren-Barring and Grahn (1995) and took a triangular approach, consisting of a postal questionnaire sent to a random sample of organizations in three cities and qualitative methods directed ata strategic sample ‘of these organizations. In addition to the questionnaire study. they used a combination of focus group technique and deep interviews. ‘The questionnaires were processed in GIS, Grahn etal, (2005) made deeper analyses ofthe daca collected by Grahn and Sorte (1985) and by Berggren-Barring and Grahn (1995), now with a special empha- sis on organizations within care and rehabilitation. Stahle (2005) used a triangular design similar to that of Berggren-Barring and Grahn (1995), however targeting individuals in Stockholm. Kyeta and Kaila (2005) used a web-questionnaire directed at individuals ina Finnish city, and the answers were processed in GIS. The study by van Herzele and Wiedemann (2003) was further developed in van Herzele (2005), and synthesized an extensive literature survey. Tyrvainen et al, (2007) and Maikov et al, (2008) synthesized a lit- erature survey as well. All of the above studies focused on urban contexts. Stockholms Regionplane-och trafikontor (2001) took a qualitative approach, using diary entries and deep interviews in a strategic sample of single persons living in Stockholm. Caspersen. and Olafsson (2006) also synthesized an extensive iterature survey. These two studies focused on mare rural contexts 1.5. Relationship between stress restoration and green spaces ss-induced illnesses have become a huge global problem. ‘According to the World Health Organization (WHO), mental health disorders and cardiovascular diseases are expected to be the two major contributors toillnesses in all parts of the world, with mental health disorders calculated fr all age groups and both sexes. by the year 2020 (WHO, 2008), Prolonged stress has serious and harmful effects on all vital organs, including the heart and blood vessels. During stress, our body organs react in many different ways, and if stess is sustained for an inappropriately long time without the possiblity of recovery, these reactions become dysfunctional and harmful withthe riskof causing deleterious changesto, for instance, the cardiovascular system and the neuro-hormonal systems of the body and of causing type Il diabetes, depression and infections (Aldwin, 2007; Tsigos and Chrousos, 2002). In particular, many psychiatric diseases are strongly associated with protonged and incorrect stress reactions, including schizophrenia, anxiety syn- «drome and, foremost. depression, exhaustion syndrome and fatigue syndromes (Aldwin, 2007; Tsigos and Chrousos, 2002). Hence, if people cannot find restoration from stress, their health will be affected in many ways. WHO has rated physical inactivity and stress astwoofthe major causes of death in the developed world, and con- sequently bas made stress-related diseases and overweight priority health prevention areas (WHO, 2006, 2008), sis not an illness per se, rather stress reactions are natu- ral and necessary. They are fundamentally the same reactions that helped our early ancestors survive. In the event of a perceived threat, stress reactions trigger the fight-or-flight reflexes that serve to sharpen our senses. Tis i expressed physicaly through, among other things, increased attention, increased muscle ten- sion, increased blood pressure, reduced digestive system activity, increased sweat gland production, increased pulse and increased production of adrenaline as well as hydrocortisone (Atkinson et, al, 1996), For early humankind, who lived on nature's terms, the body’s own adaptation mechanisms were suited to their purpose In today's urban communities, we seldom need to fight or flee. We experience stress instead. ‘As mentioned earliet, research has shown a telationship ‘between perception of the surrounding environment through our senses and human health (Ayres, 1983; Bucci, 2003: Kaplan, 2001; Ulrich, 1999). Today most research results’ converge, indicating a positive connection between how often or how long people stay in ‘urban parks or nature areas and restoration from stress and mental fatigue (eg. Kaplan and Kaplan, 2005; Nielsen and Hansen, 2007; Mitchell and Popham. 2008), 1.6, Hypotheses and aims ‘Our hypotheses can be summarized in the following two state- First: People perceive green spaces in terms of certain dimensions Second: Some dimensions are more important than others as regards restoring people from stress. The aims of the present study are to 1. identity and describe perceived dimensions in green urban spacesGrab, UX Sugar Landscepe and Ura Planning 96 (2010) 258-275, 237 2. identify the perceived dimensions people in general prefer; 4. identify the perceived dimensions people reporting stress pre- fer: 4 identify a combination of the perceived dimensions people reporting stress prefer. 2. Methods and materials ur ambition here has been to focus on town-dwellers' everyday situation. Our goals were to obtain information on town-dwellers background, in terms of sex, ge and socio-economic status: habits of visiting urban green spaces; preferences for certain qualities in urban open green spaces; prevalence of symptoms of stress, and to determine whether there are any statistical relationships between, the above-mentioned factors. ‘A quantitative survey in the form of a postal questionnaire with pre-coded questions was conducted, The aim was to obtain a representative picture ofthe situation of Swedish town-dwellers. Consequently, nine towns and cities were selected from the geo- graphical areas in which most Swedish people live: close t0 Stockholm, Gothenburg and Malmo. Approximately 70% of the ‘Swedish population lives in the above-mentioned areas (Statistics ‘Sweden, 2009) The questionnaire consisted of three parts. The frst part asked {questions about the respondents’ personal data, like age, sex, and profession, The second part focused on preferences for certain qual- ities in the urban green spaces. In this part, we did not include any activities at all, and took pains to avoid more general words, like ‘beautiful Inthe third part the respondents made self-estimations of their health status. All questions in the questionnaire were pre- coded, most often with multiple-choice options, however with an ‘opportunity to make individual remarks. To obtain a picture of the respondents’ health status, the {questionnaire included questions about the number of occasions per year each individual is afflicted by stress-related complaints. ‘As regards stress-triggered reactions, several different tests (eg, ‘Maslach, 2001; Nystrom and Nystrm, 1995) are based on knowl- ledge of clear symptoms of stress-triggered illnesses. An important testinthis context isSCI-93, which was designed by two physicians, ‘Nystrom and Nystrom (1995). t contains three modules concern- ing complaints due to stress, where one module deals with mental complaints, another with muscular complaints and a third with autonomous complaints. We chose seven questions focused on fatigue, headache, neck ache, backache, iritation,a feeling of being chased and stressed, and the common cold (vital infection). The answers were listed on an eight-step scale from “No, not at all" to “Yes, practically every day’ ‘The association between different complaints was examined using factor analysis (SAS Varimax, orthogonal rotation). One dis- tinguishable factor was formed by stress, irritation and fatigue These three complaints all have values over 0.5 (Stress 0.64, leri- tation 0.58 and Fatigue 0.58) and point to 2 strong factor that ‘we have interpreted as a stress level (Grahn and Stigsdotter. 2003). This relationship is supported by the work of other researchers, who have described how a general feeling of stress is followed by fatigue. which in turn is followed by irritation (Kaplan, 1990; Wahrborg, 2002). Based on the variables stress, invitation and fatigue, a new variable was constructed, called “Level of Stress’ (LS). To ensure that the three different vari- ables were weighted fairly in the new variable, the values were ‘multiplied by the principal component value, PCA according t0 the formula: (PC stress x stress) +(PC irritation x iritation) +(PC fatigue x fatigue)=15 (Manly, 1954; Morrison, 1976). This health index has been presented earlier (Grahn and Stigsdotter, 2003; Stigsdotter, 2005). The questionnaire was sentby mailto individuals ofall ages who were selected at random, Infodata, which keeps the Swedish per= sonal and address register, randomized the respondents’ addresses. In total, 2200 questionnaires were sent out and 163 letters were returned to sender, which means that 2027 questionnaires were delivered successfully. A total of 953 completed or nearly com- pleted questionnaires were returned; the response rate was thus 47%733 respondents were adults The adult respondents’ answers have been statistically analyzed using the statistical software SAS (SAS Statistics, 2002). Examina- tion of the respondents’ profile showed that the distribution of socio-demographic data is representative ofthe general situation in Sweden. This means that no statistically significant deviation existed with regard to socio-economic grouping (SES), sex or age between the received material and the material one could expect (Grahn and Stigsdotter, 2003; Statistics Sweden, 2009; Swedish Socioeconomic Classification Reports on Statistical Co-ordination, 1995), 3. Results 3.1, Factor analysis of preferred qualities Using SAS factor analysis (SAS Statistics, 2002), we examined the association between preferences for the different qualities per- ceived in urban green spaces (data presented in Appendix A). Kaiser's Measure of Sampling Adequacy (Cerny and Kaiser, 1977) yielded 0.92, indicating that the applicability of factor analysis in ur sample was high. The outcome indicates eight clearly distin uishable factors with eigenvalues > 1, The first factor has 41% ofthe {otal eigenvalues, the second 22%, The following six factors all have about 6% of the total eigenvalues. This indicates thatthe frst factor is the clearest, followed by the second! factor, and that the other six factors are more or less equally evident. The factors were then. rotated, Promax oblique rotation, and are presented individually below, 3.2, Description ofthe single factors Factor 1, tn this first factor, we find that “The urban park or turban open space has a nature quality’ is the variable with the highest factor loading, followed by ‘The urban park or urban ‘open space has a wild and untouched quality’ and There are fee growing lawns. This factor can be interpreted as compris ing an experience of the inherent force and power of nature, its <éynamic and intrinsic vitality. The experience includes a feeling of being in nature on its own conditions, which can be manifested among visitors in relaxing outdoor activities, such as lighting a fire. This relaxing atmosphere makes the visitor feel safe. All these variables point towards a Nature dimension of the green space (Table 1). Varabier ‘The urban pak or wiban open space contains ily eas O442s Grab UK Super Lendcap and Urban Planning 8 (2010) 258.275, Table? Factor analy, AS Fromat oblique rotation, Factor 2 Al variables, gels in Table ator analysis, SAS Promax on, actor 4. ll variables, qualities The urban parr urban open space 073 Tes possible te watch enternment lke apurkeencert 094 Tounans Ieispessibleto watch exhibitions osz ‘The urban pak or urban open spaces dee: 05 Ieipesible evs restaurant or simpler open-air os The urban parker urban open space contains awiderange 055 foreign plants, rzamentl plants and then plants The urban packer urban open space has the characteristic 057 ‘The urban parkorurban open space has diferent water 052, Tables Factor analysis SAS Fromax oblique rotation, Factor 3. Al variables, gels in The ban park or urban open space contansplaneand O87 welecu erase suraces luis possible wo havea prospec visas over the 07% The parker sn openspace has sacceriison grass ‘Therbanparkerurban open space fassocerfieléson 050, Factor 2, All variables in this factor contain elements of human artifacts: Fountains, statues, exotic plants, ponds, etc. The green space can be interpreted as decorated, as containing a core of human culture. Pethaps this can be understood in the following. ‘way. People first try to understand their environment in terms of nature and culture: both tempt people, and both dimensions are attached to history, myths and the living conditions of human, beings. We interpret the factor as a Culture dimension (Table 2), Factor 3, The two first variables in this factor are ‘The urban park orurban openspace contains plane and wel-cut grass surfaces'and ‘tis possible to have aprospect, vistas over the surroundings: These are followed by several variables with a content of open fields, pri- marily well-cut grass lawns. We conclude that this factor shows a Prospect dimension (Table 3), Factor 4. This factor contains along list of variables, and those with the highest factor loadings all concern amusements. We inter- pret this factor as follows: People are like invited guests to 2 festivity, where they can eat and érink, watch entertainments and watch other people. Everything is prepared, so they do not have to exert themselves. There are good paths, well it up, and itis easy to find restrooms and benches. We interpret this factor as revealing a Social dimension (Table 4), Factor 5. The most important variable in this factor is that the ‘green space is experienced as spacious and free. It must have a Certain quality of connectedness, so one is not disturbed by too many roads and paths. We interpret and define cis factor as aSpace dimension (Table 5). Factor 6. This factor has only three variables. however very, strong. They all concern the importance of experiencing many species: birds, butterflies, flowers, etc. We suggest that this factor concerns finding a wide range of expressions of life: The dimension is called Rich in species (Table 6), restaurant in he urban prkor he urban open space Ieispesible to shop in market stalls, ose. te o78 There ae plenty ofpeople and movementsinthe urban O74 Dak or urban open space Dathe mace of ravel “Tete ae sunny places ost ‘The urban patio eiban open space contaassevealseats 050 and benches Infelsste spending sine inthe urban parker urban open 0.44 space There ae table and benenes oe The urban parker urban open space contain oadsand_—033, pats with hard suraces, ike asphalt concrete bricks, Varabies Firiosding The urban park or rban open space experienced one Spacious and ee Ieigpocible ta Sine areas rot crosedby roa and paths 087 The urban park urban open space has lt of ees ose urban green open spaces felted 9 Warabien Factor 7.The most important variable inthis factor is “The park or urban open space contains many bushes’. followed by variables concerning play: ‘The parkcor utban open space keeps animals that children and adults may feed and pet’ and ‘There is play equip ‘ment, like swings, slides, etc” and variables related to watching active and playing people. Moreover, iis important to feel safe:'It feels safe spending time in the srban park or urban open space’ We interpret this factor as a shelter or asylum, describing a place enclosed by bushes and higher vegetation, where people can feel safe, play or simply watch other people being active, We interpret this dimension as a Refuge (Table 7) Factor. The variables in this factor all concern being in an undlis- tured environment: Silent and calm, not too many people, no noise, no liter Its important not to be startled, for instance byGra, UX Suigdater/Lanscepe nd Ura lanning 4 (2010) 258-275, 280 Table? acter analysis, SAS Promax slque rotator actor 7 Al aisle, ql: in ‘Tabte10 SAS General Linear Model, 1ype Lode: dependent varable (5 =preferences for The pak oFufban open pace contains many bushes 035 ‘Serene 493 7008 The pater uban open space Kees anunalsthat children O87 Space 320 S007 adults may eed and pet Nature eos S005 There ae sandpts ox chm species 12 Ns Theres play egipment ke swings ses, ars Refuge sez S001 Ieisposible ce wath ether people being active. plaving, 058 ature on ine rating sors, ee Prospect, os Ns Variables Fito odin The urban pak or urban pen spaces sllent an alin ry There af no ies inthe urban ark urban open space O88 (One sable to spend ime inthe wan parker ban open 084 Therese plenty of people ane movementsinthe idan 078 Dak orsrban open space “imensions of urban geen spaces, The eg pareve’ ten ea people on bikes. We interpret this factor as indicating to a retreat, haven, almost a holy place where you feel safe, We call this dimen- sion Serene (Table 8), 3.3. People’s preferences for the eight perceived sensory dimensions ‘To discover which one of the eight perceived sensory dimen- sions people prefer when they visit urban green spaces, an analysis, of arithmetic means was conducted, Table 9 shows that the infor ‘mants have the highest preference for the perceived sensory «dimension Serene, followedby Space and Nature-Prospectfollowed by Social had the lowest preferences, and Prospect also had the highest standard deviation, 34, Connection between LS and the perceived sensory dimensions Do people who report stress, measured by Level of stress (15), differ concerning theit preferences for the perceived sen- sory dimensions? In order to detect the connections between LS and each of the eight perceived dimensions respectively, a SAS Pearson Correlation analysis was conducted (data not shown) Pos- itive correlations were found between seven ofthe eight perceived rable 1 SAS Genera Linear Moge. Type itl Medel: dependenc vale L~ preferences fr the perce sensory dimensions of urban geen spaces Model: p00, N= 588 Hic a species 236 ‘S008 eure m7 2001 Socal oe Sos dimensions and LS, Only Social had a negative correlation with Ls, although not significant. Four ofthe positive correlations between 1S and the perceived dimensions were significant: Serene (p <0.05), Space (p<0.01), Nature (p<0.01) and Refuge (p <0.01), In order to investigate how the perceived sensory dimensions together affect LS, a SAS General Linear Model Procedure, Type | and Type Il, was performed. Table 10 shows that the total model 's significant. The single most significant and positive connections exist between the following perceived dimensions and LS, inorder: Refuge. Nature and Serene. Space has a clear tendency (p <0.10}, and the connection between LS and Social is negative (however ot significant). The Type i! analysis, presented in Table 11, shows the result When each of the dimensions is entered last into the model. The table shows that Refuge has the strongest and most significant positive connection with 1S, Moreover, we find that Nature has significant and positive connection. The previously found sig- hificances between LS and Serene and Space are not seen in the ‘Type Ill analysis. The properties significant for 15 in the dimen- sions Serene and Space may be clearer in the dimensions Refuge and Nature, which could explain why they no longer relat signifi cantly. On the other hand, both Rich in Species (positive) and Social (negative) have a stronger relation to LS (p-<0.10), The properties ofimportance for LS are perhaps more unique in these dimensions. Finally, a SASR-Square regression analysis was cartied out, using 1s as the dependent variable. This analysis makes independent regression models, stepwise, using Mallows C(p) to fine the optimal model, when Mallows C(p) has its lowest value (Mallows, 1973), All eight dimensions are put into the R-Square model, Table 12 shows that the single best model includes Nature, and the best two-variable model includes Nature and Kefuge. C{p) shows that including Nature, Refuge, Rich in Species and Social results in the best model, 4, Discussion 4.1. First aim: to identify and describe perceived dimensions in -zreen urban spaces The first aim of the present study concemed identifying and describing perceived dimensions in urban green spaces. We identi-2 Gra UK Super Lenscope and Urban Planning (2010) 258.275 table 2 SAS Square Stepwise Reresion Procedure, Mallow'scoefcient Cp fo finding ‘the apial mode Lethe ewes Cp Mode: dependent variable = preferences {or the prcevve reneory mensions of urban geen spaces N= 86, Model el ‘Vaables in model 7 724 Nature 3 2a 4 16 Secil nes) Richi species Refuge Nave fied eight perceived dimensions: Nature, Culture, Prospect, Social, Space, Rich in Species, Refuge and Serene, There are similarities Detween our findings and earlier research results; some of these similarities are described above (see Section 1.4) "Nature: We interpreted this factor as including an experience of the inherent force and power of nature, designed and manifested ‘on nature's own terms, Many preference studies of outdoor recre- ation environments, conducted by, eg. Kaplan and Kaplan (1989) ‘and Herzog (1987), have found thatthe presence of nature is per~ haps the most essential experienced dimension of urban green spaces: the degree to which ‘nature’ is perceived in contrast to ‘non-nature’ This dimension has been identified in many studies and labeled ‘Impression of nature’ (Grahn and Sorte, 1985); ‘Wild’ (Berggren-Barring and Grahn, 1995; Gralin et al, 2005: Maikov et al, 2008); ‘Close to nature’ (Kytta and Kahila, 2005); “Experienc- ing nature’ (Stable, 2005); ‘Unspoiled nature}mystery’(Stockholms Regionplane-och trafikkontor, 2001); "Unspoiled and adventur- ‘ous nature’ (Caspersen and Olafsson, 2006), ‘Valuable nature site’ (Tyrvainen et al, 2007), and ‘Nature’ (van Herzele, 2005) Culture: This factor is interpreted as containing an essence of human culture. This can be explained in terms of people's need {to understand the surrounding environment in terms of nature oF ‘culture. Tuan (1977, 1990), in his dialectical perspective on how people form their values of the environment, distinguished differ- tent types of environmental experience depending on the cultural and natural context, where myths and symbols play a major role, In his work Dominance and Affection (Tuan, 1984), he claimed that people creatively distort nature through cultural elements, like fountains, ponds and omamental plants. According to him, this makes the whole environment a kind of pet, satisfying our feelings ‘of power, on the one hand, and of caring for and nurturing a piece of landscape, on the other (Tuan. 1984). This dimension has also ‘been found in many studies and labeled ‘Ornamental characteris- tic (Grahn ane Sorte, 1985) ‘Culture’ (Berggren-Barring and Grahn, 1995; Grahn et al, 2008; Maikav et al, 2008); ‘Culture and hi tory’ (Caspersen and Olafsson, 2006; van Herzele, 2008; Stockholms Regionplane-ach trafikkontor. 2001; Tyrvainen et al, 2007) and ‘Gardening and splendor’. including a Magnificent display offlow- cers (Stahle, 2005}, Prospect: The variables in this factor could be summarized as, having a content of open and plane areas with a prospect, ie. vistas over the surroundings. Appleton (1975) claimed that during evo- lution our ancestors tried to find certain environments where they could settle. One of the most preferred dimensions, and clearly a necessary one, was Prospect: an open area with a view. People, Appleton claimed, instinctually prefer environments that promote ‘survival. One ofthe most important elements is having visual con- ‘trol over the environment, which allows us to detect dangers. Ulrich (1983), Orians and Heerwagen (1982) and Coss and Moore (2002), among others, have linked this type of preference to the Savannah, hypothesis, which claims that the first humans came froma savan- nah environment. This dimension has been mentioned in several studies and labeled ‘Common’ (Berggren-Barring and Grahn, 1985; Maikovetal, 2008; stigsdotter and Grahn, 2003); Prospect’(Stahle, 2005). ‘View and open landscape’ (Caspersen and Olafsson, 2006), ‘Freedom’ (Stockholms Regionplane-och trafikkontor, 2001) and ‘Sports-activities characteristic’ (Grab and Sorte, 1985) “Oppor- tunities for activities’ (Tyrvainen et al, 2007) connected to open. sreen areas. Social: We interpret this factor as an environment that is equipped for social activities. Gehl (1987) and Gebl and Gemzae (1995) have claimed that people are mostly interested in other peo- ple. Moteover, they have argued that itis of utmost importance {0 a sustainable, successful city to offer good places, easily acces sible, where people can meet, amuse themselves and watch one another. Like the dimension Nature, this dimension has been found in many studies and labeled ‘Festive’ (Berggren-Batring and Grahn, 1995; Grahn etal, 2005; Maikov etal, 2008; Scahle, 2005), Service and being together’ (Caspersen and Olafsson, 2006; Stockholms Regionplane-och trafikkontor, 2001), “Good services” (Kytta and Kahila, 2005), Town park, people's park characteristic’ (Grahn and Sorte, 1985) and ‘Facilities’ (van Herzele, 2005). Space: This actor is understood as a green environment, expe- rienced as spacious and fee and having a certain amount of connectedness. Stephen Kaplan (1990, p. 13) wrote about extent, a quality important in recreational environments: "[..]restorarive settings are often described as ‘being in a whole different world. Two properties are important to this experience: connectedness and scope: together they define what I mean by extent. Scope requires that the environments experienced as large enough that one cant move around init without having to be careful bout going beyond the limits ofthe ‘model that one is running. To have connectedness, the various parts of the environment must be perceived as belonging fo a larger whole, Without that, one must repeatedly expend effort to find the model that {is appropriate to the current momentary situation’. This dimension is very clear, has been mentioned in many studies and labeled Space’ (Berggren-Barring and Grahn, 1995; van Herzele, 2005; Maikov et al, 2008; Stigsdotter and Grabn, 2003). Spacious’ (Kytta and Kabila, 2008), Space an¢ freedom’ (Tyrviinen etal, 2007), and Feeling of being inside a forest’ (Caspersen and Olafsson, 2006; Stockholms Regionplane-och trafikkontor, 2001). Rich in species: This factor solely comprises variables demon stating the importance of finding a wide range of expressions of life: many birds, butterflies, lowers, etc. In their Biophilia hypothesis, Wilson (1984) and Kellert and Wilson (1993) have suggested that people have a strong and inherent interest in finding signs of life in their environment, as manifested by dif ferent species. This dimension has been found in other studies and labeled "Rich in Species’ (Berggren-Bérting and Grahn, 1995; Grahn et al, 2005; Gyllin and Grahn, 2005; Maikov et al, 2008; Stockholms Regionplane-och trafikkontor, 2001), and ‘Richness in nature’ (Caspersen and Olafsson, 2006). Refuge: We understood this factor as an enclosed and safe envi- ronment, where people can play or watch other people being active Appleton (1975) suggested that during evolution our ancestors tried to find a safe area where they could hide. This quality. in addition to Prospect, was a dimension of utmost importance in environments where people settled. Appleton called this dimen- sion Refuge (Appleton, 1975) The dimension is related to other results, variously labeled “Pleasure Garden’ (Grahn et al, 2005; Maikoy et al. 2008), ‘Play-Inspiring characteristic’ (Berggren Barring and Grahn, 1995), Play-activities characteristic (Grahn and Sorte, 1985), and Green oasis’ (Stable, 2005), while Kytta and Kabila (2005) found two factors: ‘Cozy’ and ‘Child friendly’ Together they an be interpreted as encompassing the same phenomenaGra, UX Sugar Landscepe nd Ubu Planning 4 (2010) 258-275, am Tabte13 Stary of esearch epertsMentiving an desrbing experience dimensions in green environments silat those found inthe pesent tu. Reseach repons Teipenenced sensory cmensons nthe preset aay Berggren-Bicring and Gran (1995) x x x x x x x x Stockholms Reglonplane-och waflenter(2001) x x x x x Grab et (2005) x x x x x x x x van Herzl (2005) x x x x x ° x x x x x x (Caspersen and Olatsson (2006) x x x x x x Tyevanen et (2007) x x x x x ao et (2008) x x x x x x x x Serene: This factor is about being in an undisturbed, silent and calm environment, which can be interpreted as an environment for retreat~a virtually holy and safe place. Jensen (1998) found that {quietness was one ofthe primary dimensions motivating people to Visit green spaces. This dimension has been found in several stud- ies and labeled ‘Peacefulness'(Berggren-Barring and Grahn, 1995; Grahn and Sorte, 1985; Grahn et al, 2005), ‘Serene’ (Maikov etal 2008), Quietness’ (van Herzele, 2005) and ‘Calm’ (Stahle. 2005), Tyrvainen and her associates (Tyrvainen et al, 2007) found a pair cof factors: One positive -'Peace and tranquility, and one negative - NNoisiness, while Kyttd and Kahila(2005) found two actors: Peace- ful vs, Restess, noisy’ and ‘Tidy vs. Untidy’ Together they seem t0 capture the same phenomena, Consequently, our findings are in line with previous results Despite the different methods used (qualitative. quantitative of literature surveys), the different target groups addressed (organi zations or individuals) and the different contexts studied (urban green spaces or countryside), the results are very similar. The dimensions have been identified and described one by one, often ‘many decades ago, Moreover, many researchers have suggested ‘that experienced qualities in green spaces can be divided into cer- tain classes (see Table 13). As we understand them, these qualities have been detected and described in a way that makes it possible {or us to compare them with our results: van Herzele (2005) found five dimensions - all of them detected in the present study. Grahn and Sorte (1985) identified six dimensions all of them discovered here as well. Stockholms Regionplane-och trafikkontor (2001) and Caspersen and Olafsson (2006) have mentioned seven dimensions six of them detected in the present study. What is interesting is that Stockholms Regionplane-och trafikkontor (2001 claimed that Serene’ is of the utmost value, however so crucial and important ‘that is not necessary to bring it up as an eighth dimension in plan- ning. Berggren-Barring and Grahn (1995), Grahn et al (2005) and Maikov et al. (2008) have identified eight dimensions-all of them found in the present study, yet in two of the studies (Berggren- Barring and Grahn, 1995; Grahn etal, 2005) they were found more or less in the opposite order: The Serene dimension was their first factor and the Nature dimension their last, Kytté and Kabila (2005) found nine positive and nine negative dimensions, Most of ‘them Were paits of dimensions such as ‘Peaceful-Restless, noisy’ However, in their further analyses, only the positive dimensions ‘were used. Seven of these dimensions have been found here as well. Tyrvainen and her associates found eleven dimensions. Eight dimensions were positive and three negative. Five of the positive dimensions and one of the negative dimensions were found here to0. Lastly, Stahle (2005) found ten dimensions. Six of them were discovered in the present study. Stockholms Regionplane-och trafikkontor (2001), Stable (2005), Kytta and Kahila (2005), Tyrvainen et al. (2007) and Caspersen and Olafsson (2005) have identified dimensions not found here Most of them are activity classes, as they have mixed activi- ties with qualities connected to the environment, Stable (2005) found four activity lasses: ‘Pieni activities’, Water activities (e. playing in water, bathing, swimming), Cultivating activities (eg. ‘growing vegetables, keeping small animals like rabbits and guinea pigs) and “Activities connected to big green commons’ (eg. soc cer, temporary events) Stockholms Regionplane-och trafikkontor (2001) and Caspersen and Olafsson (2006) have a seventh factor: ‘Sports-activites characteristic. Inthe present study, activities are completely separated from qualities attached to the environment. ‘Tyrvainen et al. (2007) found the factor ‘Forest feeling’—that the area feels like a “real” forest, We have focused out study on urban sgreen spaces. In spite of this, a forest feeling could be related to ur factor ‘Nature’. however being more of an identity dimen- sion. Tyrvainen and her associates (Tyrvainen et al, 2007) also found two factors related to security: ‘Scariness’- that the area feels dangerous or threatening, and ‘Unpleasantness' ~ that the area is neglected, abused or damaged. Kytti and Kahila (2005) found a factor related to security as well: Secure’, which we have found to be an important property in several dimensions, how- lever most obvious in Refuge. Furthermore, Kytta anc Kahila (2005) found the dimension ‘Good connections’, which concerns traveling. to the site. Tyrvainen and her associates found two factors con- cerning aesthetics: ‘Beautiful landscape’ and ‘Attractive park’—the latter likewise connected to exceptional beauty. In the present study we have avoided all words and concepts concerning plain aesthetics. [Nature isa characteristic found in all nine studies, including this, ‘one, Socialis lacking only in the study by Tyrvainen etal. (2007) and Culture is lacking only in the study by Kytts and Kahila (2005). In studies relating outdoor life and preferences to green spaces, amore anthropocentric perspective on green spaces is needed, Nature is ‘what distinguishes green spaces from the man-made landscape, Social is one of the most important qualities of green spaces for people, and Culture is revealed in how the green space is designed, land managed, This may explain why these thtee dimensions are distinguished from the others. 42. Second aim: identify the perceived dimensions people in general prefer Curresults show that Serene was the most preferred dimension, followed by Space and Nature. Rich in species and Refuge were found in the middle, while Culture, Prospect and Social were the least preferred, Jensen and Koch (1997) and Tyrvainen etal. (2007) found that ‘Peacefulness' was the most valued quality among the Visitors they asked. Peacefulness does not mean that the area is absolutely quiet, you can listen to the wind inthe trees and you can, listen to bitds, but you are not affected by noise. Berggren-Barring and Grahn (1995) found thatthe dimensions Serene. Space’ ‘Wild and ‘Rich in Species’ were connected to the largest urban green spaces in the cities they investigated.m Gra UK Supt Lendscape and Urban inning 8 (2010) 268-275 Kaplan (1990) claimed that one of the most important quali ties for a recreation area is its compatibility. That is, what you as 2 visitor want to find and want to dois exactly what the area is suit able for and allows you do. According to Gibson's (1978) theory, the area affords you the right opportunities. Jensen (1998) found five dimensions of motivation or push-factors that make people ‘want to go out in nature areas, and these could be linked to both ‘compatibility and affordances. These were ‘Experiencing nature’ such as finding areas with mystery and wildness, coming in close ‘contact with different species, enjoying views ofthe landscape and enjoying sounds and smells from nature. n our study, the dimen- sions Nature, Rich in species, Prospect, Serene and Refuge might together satisfy this need. ‘Another dimension of motivation in Jensen's (1998) study was Peace and quiet'—finding 2 place where it is quiet, a place far away from densely built-up and populated areas, finding a change {rom everyday city environments. The present dimensions Serene and Space can together satisfy this need. A third dimension of motivation was ‘Social stimulation’, such as meeting new people, talking to a variety of new people, showing that you dare to do something, and being with friends. In our study, the dimension Social can satisfy this need. The two last dimensions of motiva- tion were ‘Physical exercise (eg taking a walk} and ‘Strengthening family ties’ (being in places together with my family): these are not clearly connected to any of the dimensions found in our study, From the above, we can see that Serene satisfies several needs. You are allowed to experience different sounds in nature, and are not affected by noise and crowds af people. 43. Third aim: identify the perceived dimensions people reporting stress prefer ‘As stated earlier, in studies relating outdoor life and prefer- ‘ences to green spaces, a more anthropocenttic perspective on green spaces is needed, However, for people experiencing stress, dimen- sions other than Social and Culture are more preferred, such a5 Refuge, Nature and Serene. This could be depicted as toning down the anthropocentrie perspective, allowing a more empathic and participating perspective to appear for individuals experiencing. stress, Such individuals need to be involved on a more basic level inthe green space. The eight dimensions discovered and described here consist of messages that manifest themselves through many different sensations, perceived through our vision. hearing, sense ‘of smell, locomotion, ete. Hence, the dimensions should be able to ‘communicate with many aspects ofthe individual's body. muscles, ‘thoughts and feelings, and thus the individual's preferences and decisions to visit the area or not Earlier research has shown that the most ill nd vulnerable indi viduals ~ those striving to Gnd balance within them ~ seem to be ‘mast sensitive to the perceived dimensions in the environment (Ottosson and Grahn.2008,2008; Ulrich, 1999). The sensory experi- ‘ences of outdoor areas (green areas as well as moreuilt areas) may be of the utmost importance to whether or not stress-reduction effects are felt. Our results show that individuals reporting high levels of stress have preferences for the dimensions Refuge and Nature. To a certain degree, Rich in Species, Serene and Space are important, while Socal isa dimension that could be interpreted as adding to the total stress burden. A person who is affected by stress finds it increasingly difficult to understand, sympathize with and tolerate other people (Whrborg, 2002). Ulrich (1999) pointed out that people have an inherent capacity to interpret dimensions in nature in terms of stress reactions. We easily adapt our stress reac- tions to appropriate levels in nature, while urban areas are more difficult to interpret through our reflexes. Kaplan (1990) talked bout the importance of restoring directed attention, where nature is the optimal place for using involuntary attention and resting directed attention, According to Kaplan (1990), characteristics t are good for mentally fatigued people are an impression of natu (the dimension Nature), the experience of extent (the dimension Space) and the experience of fascination (the dimension Rich in Species) ‘Ayres (1983), Bucci (2003) and Ehrenzweig (2000) have all emphasized the importance of sensuous experiences, where ‘unconscious information can be stored. The Refuge is a dimension that involves opportunities, or affordances, to be in 2 small safe place, where you can be alone, come close to the vegetation, and Ihave possibilities to use all your senses. 44, Fourth aim: identify a combination ofthe perceived dimensions people reporting stress prefer ‘Our results show that if we want to create an area with one single dimension, Nature would be the optimal solution, However, allows coefficient shows that a combination of Nature and Refuge ‘would be a better solution, especially if there are no signs of the dimension Social at all. Our interpretation is that an area offering, ‘these dimensions would be more preferred, because it offers better possibilities for restoring various capacities. In the Type Ill analysis, Nature and Refuge were the only «dimensions with a positive significance. The characteristics most preferred by people reporting the highest levels of stress are also found in the dimensions Serene and Space, but these certain char- acteristics are probably even stronger in the dimensions Refuge and Nature, which causes Serene and Space to disappear in the Type Ill analysis. Basically, experienced qualities in urban green spaces should be able to meet the individual's andjor the group's vital needs, ie. beliefs, desizes, emotions, etc, Individuals who report stress are most sensitive to the environment, and hence most difficult to consider. Based on the theories above, we learn that we perceive our environment through all our senses and that the informa- tion is processed and stored at the subsymboli, symbolic imagery and symbolic verbal levels (Bucci, 2003). People suffering from stress have difficulties interpreting their environment, and dis- sociation between the senses, emotions and cognition can occur (Bucci, 2003), Moreover, those suffering from stress instinctively seek out sheltered areas, and do not want tobe social (Ulrich, 1998; ‘Wabrborg, 2002), We interpret the present results as showing that people who report high levels of stress need peaceful, more undif ferentiated areas where they have a chance to make contact with, their basic senses and more deeply stored subsymbolic informa ton, Because dimensions and characteristics in our environment are interpreted immediately through cognition and emotions and ‘through our pre-linguistc vitality affects (Stern, 1985, 1993). sym- Dols from nature have a chance to spark creative processes that are important in the restoration process (Bucci, 2003; Ottosson and Gralin, 2008; Searles, 1960), Rich in Species entails the pres- fence of fascinating objects, which Kaplan (1980) found to be crucial to restoring an individual affected by mental fatigue. A combina tion of the dimensions Refuge, Nature and Rich in Species, wh the dimension Social is toned down, could be interpreted as being ‘most preferred by people reporting the highest levels of stress and thereby such a combination may offer the optimal place for recovery, 5. Conclusions and future perspe Historically, public health has always been an important per- spective in city planning. Thanks to developments in modern ‘medicine as well as improved living standards and living environ- ments with, eg. better hygiene, several communicable diseasesP. Gra, UX Sugar Lanscepe an Ubu Planning 96 (2010) 258-275 2m have nearly disappeared (Puranen, 1984). However, other threats to public health have developed (WHO, 2006, 2008). People in the West are living their lives farther and farther from nature. The majority of Swedes now live in cities and, moreover, spend most of their time indoors (Qvarsell and Torell, 2001). This trend has also been observed in other developed counties all over the world, nnot least in the US (Pergams and Zaradic, 2006), This and other ‘rends in the developed world are connected to a more sedentary and stressful life (Pergams and Zatadic, 2006; Wabrborg, 2002}, ‘Today, when recreational areas are analyzed and described, for instance atthe prospect of exploitation, in Sweden, and presumably in many other countries, its clear that decision-makers have not, been given a fair chance to assess the importance of such areas to the outdoor life ofthe town population (Grahn et al. 2005; Stable 2008). The qualities in urban green areas are most often not appro- priately accounted for in the maps and documents that form the bass for diferent kinds of decisions When urban green spaces ate described, no values are presented that refer to people's outdoor life, their preferences, needs or health (Grahn et al, 2005; Stahle 2005). The present findings suggest that preferred urban green areas play specific roles and embrace qualities of clear importance to health and outdoor life. In conducting the present study, our goal has been to deepen ‘our understanding of why certain urban green spaces may be con- nected to mental health. This concerns how humans perceive and process sensory information, andit also concerns preferences. Here, ‘we have investigated individual preferences using a representative sample of the Swedish adult population, The results identify and describe eight perceived sensory dimen- sions in green urban spaces: Nature, Culture, Prospect, Social Space, Rich in species, Refuge, and Serene. People in general pre- fer the dimension Serene, followed by Space, Nature, Rich in Species, Refuge, Culture, Prospect and finally Social, ranked in order. However. the dimensions Refuge and Nature are most strongly cortelated to highly stvessed individuals’ preferences, indicating a need to find the most restorative environments. A combination of Refuge, Nature and Rich in Species, and a low or no presence of Social, is the most preferred urban green space, and could be interpreted as the most restorative environment for stressed indi- viduals, Inthe present debate on the dense city contra the sprawied city, the dense city is often considered as a suitable direction for future city planning. the idea being that such cities promote sustainable development (Hardy, 2004). Yet the health-promoting perspective ‘on urban green spaces also concerns the sustainable development of cities, New building within city limits and densification of the city may result in the thinning out and disappearance of the very value of parks and other natural and recreational areas, because the perceived dimensions appreciated by people are closely tied to factors such as quietness, size and shape. Some of the most important perceived dimensions for people reporting high levels of stress, and thereby for promotion of health, are also the most sensitive (0 disruption, because these dimensions ~ Serene, Space and Rich in Species ~ require large land areas (Berggren-Baring and Grahn, 1995), Thus, these dimensions are thinned out and thereby disrupted, the health-promoting qualities of parks and other natt- ral and recreational areas may be lost. The results presented here should be considered with a view to future urban planning, where turban green spaces can be seen as a resource of importance to public mental health. However, before the eight perceived sen- sory dimensions can be used by practitioners as tools to promote health through design and urban planning, more researchis needed to understand these dimensions, and how they can be detected, developed and maintained. ‘Acknowledgements This paper was made possible by financial support from ‘The Swedish Research Council Formas. We would like to thank our statistician, Associate Professor Dr. Jan-Eric Englund, Swedish University of Agricultural Sciences, Alnarp, for his kind assistance,Appendix A. Gra UK Supt Lenscape and Urban Panning 9 (2010) 268.275 Factor analysis, SAS Promax oblique rotation, of dimensions in green spaces. Rotated Factor Pattern. N=G43 (adults only). Loadings of each ‘of the scored items, forming the basis of analysis (loadings <+0,30 not shown). Possbletolghta fre 230 Foveiga plants {Gy park eharaense Pond, ana Blowers Wooded pastaregualty Plane. wel-cut grass Got laine Special pak annals General good iting Ross welitup Paces sherere trom the wing Sunny places Shady places Several ets and benches Tables and benches Path wth har surfaces Experienced a spacious reer net rosea by paths Sent and clin Net crowded Plenty of pepe Nemopeds Watching people being active ean se well maintained Variance explained bythe factor ignerngetherfacors 562710 os O76 07% 089 087 ass sie asso References Tere New York * Appleton. |. 1975. The Experience of Landscape jn Wiley, London, ‘ason RL. Atkinson, RC. sath, EE- Bem, DJ. Nolei-Hoeksema, S. 1996, ‘ids iatoducon eyelet ace Calege Palos Fore ‘Ayes JA 1983, Sensory Iteration and the Child Western Paycolaricl Services, sel. 1955 Landscape Pate. Perception and Process, EREN Spon, New York ‘an den Seg AE. Kool, SL, van der Wp. NY. 2002. Envonmettal preference tnd restcatin iowa they related?) Eon Psychol 22, 125-146, Betgques-Buting A-M, Grabn P. 1995. Grnstrultuens beryl fr anvind tinge (the sigacance of the glen structure fer peoples se) Pablsed ‘Sciences, Department af Landscape Planning Uituna in Swedish English sim sj. Albin MA rah. jcabsson, H. Aeds J. Wadbre J. et a. 2008, ‘Recreatlena vives ofthe nituralenvosment in elation to eghbourod ‘sac, pal ebay and welling} Fpidenia Cnn ce, Wi, 2002, Varieties of earociatve expenences.Paychoana dA Lane, 5) Muy, 2008 Soy atertn These aneGra, UX Sugar Lanscepe and Ubu Planning 4 (2010) 258-275, 2s Caspesen, O41, Olson, AS, 2006. Opleveseverdier of det grenne Ninoy [Espevencing vate: i the green sttctre of Copenhagen, Research Report. “analytic correlation matrices. Mulivar Behav. es. 1243-47. ~ we SK, 1968 Lanéseape survey and analysts ob sation bai Plann Outeek ‘preschool een fea Payee Ma, 161-202 hrenaweig, A 2000 The Hidden Order ofA A Study ote Psychoeny ot Arise Imazinaton Phoenix London este). Otefhalm, Al, 1973, Lanéskapsaraker Vurdeting av faut og Testeibk hiaie’#natlandshapet [taneseape characters stinaling of “Snderstanding and aevthetielquales inthe natural landscape], Researeh Report. The Agel University a Norway Ac in Nerwerian. ‘cen 1987 Lie between Sulings’ Using Pubic Space Van Nos Geb), Gebawe, L195, Pble Spaces and Publi fe Danish Arteta ress ‘ibe |}. 1979 The Fealogcal Apprasch te Visual Fereption Houghton Mit, an, 191, Landscapes in our minds: people's chee of recreate places in aha, P. stigséotter, UK. Bergeren-Raring, AML, 2005. Human issues. tn ‘Wergus, AC. Duemb. Unchelm,c. Opperia, B. Pale. S. Taig 5. Gran Sorte 1985 tar snvandsparken? Om arganiserade grupprs br ‘Wgronomriden [How are utban pare used? Thee of whan een spaces “iterene groups of eranizatios} Resear Report The Swedish Univers of [Agncututal ences, Alnarp la Swedish Engish Suita) Human Seclements, Nairobi . ary, 2008 Renassns fr den tradonellataden enalssance forthe trade Marti, 7, Evans, GW, Jamner, LD. Davis, DS, Gatling, T, 2003, Tracking restoration natural ad urban ele stings | Ensvon Poy 22 108-12 Hering, TR 1967, & conntve anabyse of preference for naira environments mountain, canyons csets Landscape Res 6140-132. fea abs ly et tn ps edn eatin ‘eteinary an Agricaltirai ives lrsolmnDanish: Engh sat Jensen 5 Koch, NE 18), rity skovene1975/7)-1993/94 [outdoor rere= “on in forse 1876)77-1993)9) Research Repos 20, Forest Landscape, The ambiege ° Kaplan 5 1880. Parks for he future psychologist view In: Sot.) (6). Fars emit fare Ie sedative Agel Scenes Alp. 2-22 context of eat ature, In Bale P(E Urban Place Reconectng With ‘he Natural orld Ms Prevs,cambricge, EA pp-271-298, yt M, Kaha, 2005-The perceived quay fetrs ofthe environment anetheir ‘concent aces Ins cal Cth, (| Forest, Tees Homan Health SndWel-beine Media & ScenticPubehers Thessaly p. 337-951 allows CL. 1873, Some comments on Cip.Technomettics 15. 661-675. Manly BF). 1994 Alevanate seus Methods A Primer. second e@- Chapman ‘bial tendon. ‘yet ibanang Bunt The Coo ng Nae 8 el, ti "Michell, Pophar, 2008 fect of exposure co natura envionment on health reese anebeeanapopulton ay. Lact 32135-1050." Nieleen TS, Hansen KB, 2007, Do gren areas affect helt? Resa rom Danish Survey onthe se of eenaieas ana heal ineators Health Pace 13, 838-850, sn Hyams and etn = 30-3. A proce rai, Ontenten, E, Lune, 1, Seva Ravenel. 1, 1974, Forventninger ti reizesionsmlee 1 Osomatka [expectations of recreation emroammest 0 sierra) esearch Report Norwegian Institute for Urban abd Regional a 1d Reinho, ans 6; Meagan. 192- Eee pies fancy Stn ‘weston. Cran, P2005, A omparizon else time spent na arden with Tesure me spent indops: on measures of restoration nese In geate re Landseape Res 3025-55, oven ea 208 Th le tnt ein in ci haiti Hn Penning Ronse. 1973 Aernativeapproaiestolandscape appraisal anéeval- ation, Planning esearch croup Report 11. Midalesex Pelyeehnic Enel espns ORM. Zarate PA. 2006 lave of ature USbvcening veal Sve pinnae caer aes ope)Evon Maas. 8 sla Hetetine RF, Goodman, P3970. estat Therapy Excitement and Grow ‘nthe Humas Personality ita Book Hew York, "750-1940 [ubecolars An less acertence and ft cause Sweden 1750-t940 Publis Dissertation, Ummes Univers, Uines in Smear English summary). ‘vatsel Tore U, 2001, Humanist hilsoforsknng Et vixandeforiningstl. in: Qrase R Tol (eds) Humantaktaetershning-enforkaingsover- ke Stadentiterst Lund pp 3-22 (in Swed). Searles HE, 1950, The Nonhuman Environment in Normal Development and in Shafer, EL, Hamilton, [F Schmide. EA, 1869 Natural lanéseape preferences: a predictive model Lene Ree 11-18. Stanford neycopedi on Posopy, 2008 stanford University Stanford, CA Stata veurbookof Sweden 2008 Ocal statistic of sweden, Statist Seen, ‘Stokaoim The Perceive Self-Ecologiea and Interpersonal Sources of Ser-knewedze anoige Uieriy esc Ma pp.atais. suds at ein Mere pm iv ey iit schael of Architecture Stckolm in swedsh: Engst summary) sditet. U, 2003. Landscape archiecute and hele Evicence-se Death Suffering fom bueno eases |. Therap Morte 1438-48 helms Reglonplane-ech taflaonter, 2001, Upplevelseviréen._Socila the elon gien st] Reorplneoch luda, Stk ‘Swedish Socio-economic Classification Reports on Statistical C-erination, 1995, ‘Statstics Sweden Stockolm, Tues C,Chrousos CP 2002 Stes te endoplasmic reveulum ane ns cesis- Suan Wei, TOOT Space and las Te Fe fs u 7 Tuan ¥-F, 1984 Dominance and Afecion Yale University Press New Hoven, Tan, YF” 1900. Tpephil A Study of Envtsonmental Perception, Actes 2nd "yee, Lane Schipper 207 Ton raping alas of Uineh. RS” 1585 Aesthetic ad fective response fatal ewontet In: A= ‘nan. Woblwrl[F (Eds Human Behaie and Enitonment Vol 5 Behater Design Recommendations Jan Wiley Sens New York pp 27-86. Ltneh KS. 2008 Euidence-basedReale-careachtetre Lancet 268, 28-53. sei A203 Ae your doef ou nd eerie sible and attractive urban green spaces Lanéscape Urban Plan 93, 109-126. Wahrboig P2002 stress och den aya eblsan tess and the New eat Natu Urban Greening 5, 199-212 uuon ED: Tanta Hae Uaesiyecambage, ediacentefactsheets len WHO, 2008: Depression Programmes and Projects. Menta Health. Word Hea ‘iganizaton. Retrieved July 16, 2008 fom "itp ww Wout]